1. Field of the Invention
The present invention relates to imaging apparatuses. More specifically, the present invention relates to an imaging apparatus with which a picture of an improved quality can be taken of an object having both a bright region and a dark region, i.e., a high-luminance region and a low-luminance region, and that allows precise analysis of change in the luminance of a picture taken of an object whose luminance changes.
2. Description of the Related Art
Recently, digital cameras are becoming common. In a digital camera, image signals captured are digitized, and the digitized signals can be transferred to an information processing apparatus such as a personal computer (PC) via a recording medium such as a flash memory, via a cable, or by infrared communication. The personal computer, having received the data, is allowed to display a corresponding image on a monitor such as a CRT display or a liquid crystal display.
A picture taken by an imaging apparatus is composed of data of a plurality of pixels. For example, in a video camera, a picture is formed based on lights received respectively by 720×480 photoreceptor elements corresponding to an array of 720 horizontal and 480 vertical pixels. In the case of a still camera, the number of pixels is, for example, on the order of several hundred thousands to several millions. The pixels correspond one by one to the photoreceptor elements in the imaging apparatus. The value of data for each pixel (pixel value) is proportional to the amount of optical energy that is incident on the corresponding photoreceptor element during an exposure time. That is, the pixel value of each pixel is proportional to the amount of optical energy that is incident per unit time.
When eight bits are assigned to each pixel, the pixel value of each pixel takes on one of 256 values ranging from 0 to 255. 0 represents the minimum luminance, indicating a dark region for which no optical energy is received by an associated photoreceptor element. On the other hand, 256 represents the maximum luminance, indicating a most bright region. A pixel value of 2 indicates that the pixel receives twice as much optical energy per unit time as a pixel with a pixel value of 1.
As described earlier, a video camera typically has 720×480 pixels, and a still camera typically has several hundred thousand to several million pixels. When an image is taken, exposure time is common among the large number of photoreceptor elements corresponding to all the pixels.
For example, if an exposure time for a photoreceptor element corresponding to a certain pixel of an image A is a period from time T1 to time T2, imaging time is the period from time T1 to time T2 for all the pixels constituting the image A. The length of the period (T2-T1) is usually on the order of 1/30 to 1/1,000 seconds. The period (T2-T1) is referred to as an “imaging interval” for capturing the image A.
For example, a motion picture is captured by generating one image, i.e., one frame, at a cycle on the order of 1/30 to 1/1,000 seconds and successively capturing a plurality of frames. Each frame consists of a single image, and a frame image is formed by exposing all the pixels from a particular time for a particular time interval (exposure time).
As described above, in a conventional imaging apparatus, imaging is carried out by exposing all the pixels for the same exposure time. This has raised various unfavorable restrictions, which will be described below.
For example, a problem arises when an object having both a bright region and a dark region is imaged by an imaging apparatus such as a video camera or a still camera. If the exposure time of the imaging apparatus is set to be short in order to avoid an overflow of optical energy received by a photoreceptor element on which a high-luminance region of the object is projected, sufficient optical energy is not received by a photoreceptor element on which a low-luminance region is projected, so that the projection image of the low-luminance region becomes completely black.
On the other hand, if the exposure time is set to be long so that sufficient optical energy will be received by a photoreceptor element on which a dark region of the object is projected, excessive optical energy is received by a photoreceptor element on which a bright region of the object is projected, so that the projection image of the bright region becomes completely white. That is, it has been inhibited to avoid an overflow in a projection image of a bright region while preventing a projection image of a dark region from becoming completely black. This problem will be referred to as a “first shortcoming of the related art”.
Furthermore, another problem will be described below, which will be referred to as a “second shortcoming of the related art”. This problem arises, for example, when an object whose brightness is changing rapidly in time over a dark background is successively imaged. In order to analyze how the brightness of the object is rapidly changing in time based on a plurality of image frames captured, successive imaging with a short exposure time is required.
If successive imaging is carried out with a short exposure time, however, exposure time for each one of the images is short. Thus, sufficient optical energy is not received by a photoreceptor element associated with the dark background, resulting in a completely black projection image, so that a favorable image of the dark region cannot be obtained.
On the other hand, if successive imaging is carried out with a long exposure time so that sufficient optical energy will be received by a photoreceptor element on which a region of the dark background is projected, the time interval of each image frame becomes long, inhibiting precise analysis of the rapid change in the luminance of the object. As described above, when successive images of an object having both a bright region and a dark region are taken for analyzing change in the luminance of the bright region, it is not possible to obtain clear images for both the bright region and the dark region. This problem will be referred to as a “second shortcoming of the related art”.
Imaging apparatuses that overcome the “first shortcoming of the related art” have been proposed, for example, in “David Stoppa et al., “A 138 dB Dynamic Range CMOS Image Sensor with New Pixel Architecture”, IEEE International Solid-State Circuits Conference Digest of Technical Papers, pp. 40-41, 2002”, Japanese Unexamined Patent Application Publication No. 2001-326857, entitled as “Enzan kinou tsuki satsuzou soshi”, which could be translated as “Imaging device capable of calculation”, and “YOSHIMURA Shinichi, “CMOS imeeji sensaa no kougashitsuka to apurikeeshon”, Technical Report of IEICE, ICD 2001-97”, which could be translated as “Improvement in picture quality of CMOS image sensor, and applications thereof”.
Features of imaging apparatuses disclosed in these documents will be described below. Initially, an electric signal of an initial-setting electric-signal level (Vinit) (e.g., an initial potential) is set to each photoreceptor element in an imaging apparatus.
Then, exposure starts by imaging. During the exposure, the electric signal set to each photoreceptor element flows out and tends to zero. This is because when an optical energy is received by a photoreceptor element, an amount of electric signal in proportion to the amount of optical energy flows out from the photoreceptor element. The rate of the electric signal tending to zero is proportional to the amount of energy that is incident on the photoreceptor element.
The level of the electric signal, being attenuated as time passes, is compared by a comparator with a predetermined threshold value (Vth). Then, a time (Tc) when the electric-signal level of each photoreceptor element becomes lower than the threshold value (Vth) is recorded. For each photoreceptor element, a value (Vinit−Vth)/Tc is calculated using the time Tc, whereby the amount of optical energy that is incident on the photoreceptor element per unit time is known.
As described above, optical energy received by each photoreceptor element corresponding to a pixel is measured, and an appropriate time for receiving optical energy is set for each photoreceptor element, avoiding reception of excessive or insufficient optical energy. Accordingly, the imaging apparatus allows imaging over a wide dynamic range. Accordingly, an overflow is less likely to occur, so that a projection image of a dark region is prevented from becoming completely black without causing an overflow. This arrangement overcomes the “first shortcoming of the related art”, described earlier.
Even in the arrangement described above, however, with regard to a photoreceptor element for which the amount of optical energy that is incident thereon per unit time is extremely small, the electric signal thereof is hard to be attenuated, so that a problem arises that it takes a considerable length of time before the electric-signal level becomes lower than Vth. This problem will be referred to as a “third shortcoming of the related art”.
Furthermore, even in the improved imaging apparatuses, after a time when the electric-signal level becomes lower than Vth has been recorded, each photoreceptor element stays idle doing nothing until recording of a time for a photoreceptor element associated with the darkest region is executed. That is, in the imaging apparatuses, imaging interval is the same for all the pixels. Thus, the “second shortcoming of the related art” is not overcome yet.